added a research item
The Command and Control (C2) for the Smart Grid must provide the advanced computational algorithms that will enable distillation of actionable intelligence in a manner that has been demonstrated to be scalable to millions of nodes. Further, it must provide the Secure Smart Grid Common Operating Environment (S2COE) middleware layer capability and cyber-security that is needed to safely and securely manage the complex software applications operating within this System of Systems (SOS) throughout the Smart Grid (new and legacy together), while minimizing costs to the ratepayers. We propose to demonstrate in the New York City and the surrounding region, an integrated C2SOS platform that will provide the intelligence, interoperability among legacy systems as well as new and future systems, and cyber-security that is essential not only for rapid and effective deployment of a Smart Grid in both urban and suburban environs across the country.
Northeastern Infrastructure War Gaming: We believe one of the missing ingredients to the homeland defense is the need for electric grid threat simulators for war gaming – one that allows hundreds of threat scenarios to be examined on the computer and responses devised and trained for. First and foremost, the topologies of the regional high voltage grids managed by Regional Transmission Operators (RTO) and Independent System Operators (ISO) must be combined on the computer with local power grids, and more generally, distribution networks managed by utilities such as ConEd in New York City, National Grid on Long Island, PSE&G in New Jersey, and hundreds of other public and private generation and distribution companies across the country
Electric Grid Analytics Learning Machine (or EGALM) system is a machine learning based “ “brutally empirical” analysis system for use in optimizing the performance from one or more electric grid components. The EGALM system optimizes production and distribution and/or consumption of electricity while minimizing equipment failures and therefore costs. Normalized data are processed to determine clusters of correlation in multi-dimensional space to identify a machine learned ranking of importance weights for each unique attribute of the data. Predictive and prescriptive optimization on the normalized electric grid data are performed utilizing unique combinations of machine learning and statistical algorithm ensembles. The unstructured textual data are classified to correlate with optimal electrical grid performance using physically real or theoretically calculated systems to provide categorization results from labeled data sets to identify performance patterns in mean time to failure of critical electric grid equipment.
New Texas Wind and Gas-to-Wire Electricity Generation Provide capacity up to 1000 MW $393,000,000 (April 2002 Plan) Provide added capacity up to 1500 MW $247,395,000 (Dec 2002 Plan) Provide additional capacity up to 2000 MW $312,795,0003 (Dec 2002 Plan)
Susceptibility of vrious components of the NYC electric Grid in Williamsburg Network during a heat storm in 2017.
Ecomagination is a competitive force for growth across GE’s businesses. With more than $85 billion in sales and services through 2010, ecomagination is a business strategy that represents an area of continued strength for the company. GE is committed to continuing that success.
The goal of this project is to design, build and evaluate smart power grid demonstration projects in the New York City region. Demonstration projects will incorporate new sensors, extensive communication facilities, distributed control, distributed generation (e.g. solar, wind, cogeneration), storage systems, and pluggable electric vehicles.
Renewable Energy: intermittent non-dispatchable generators (ING). Loads are control by end-energy users. Independent system operator (ISO) control the generation on 2-4 second cycles and follows the load. Spinning reserve power is supplied by a number of generators that are connected to power grids to be used as needed to balance grid generation with grid load. Gas fired peaking units are used as spinning reserve. When ING sources are not available, spinning reserve units are used to balance grid generation with grid load.
The susceptibilty ranking system gives live measure of performance of each network from a feeder perspective.
PROJECT SUMMARY The use of Distributed Electric Energy Storage (DEES) for the real time support and optimization of the electric generation, transmission and distribution (GT&D) system has been limited to date to pumped hydro, primarily due to a lack of cost-effective options and/or sufficient value bases, as well as actual field experience. Recent developments in advanced energy storage technology, including a number of demonstration and commercial projects, are providing new opportunities to use energy storage in grid stabilization, grid operation support, distribution power quality, and load shifting applications. Our project team proposes to characterize the leading DEES markets for New York State, including a projection of the respective capacities and range of values. We will then drill down into the detailed cost, benefit, risk, and uncertainty benefits for use of an exemplary DEES technology, a 10 MW sodium-sulfur (NAS) super-battery in an urban substation in a critical load pocket of New York City. As a baseline, we will conduct a net present value (NPV) analyses to establish a classical benefit-cost ratio figure of merit. In parallel, we will build a template for using a more sophisticated stochastic model, Real Options, to demonstrate the advantages of the use of this advanced market analysis tool to better understand and justify market penetration for DEES in the GT&D systems of New York State. Our proposal team consists of leaders in key aspects of DEES and GT&D market analysis: the Center for Economic and Environmental Partnerships in Albany, Columbia University and Consolidated Edison Company (Con Edison) in New York City, with NGK Insulators, Ltd. (NGK) in Nagoya, Japan and Technology Insights (TI) in San Diego. NGK is the supplier of the NAS battery system and will provide related cost and performance information as well as the overall DEES experience in Japan. TI has been supporting NGK's NAS market and project development in the US, including a background of such effort with Con Edison. TI has recently completed an EPRI/DOE sponsored Handbook of Energy Storage for Transmission & Distribution Applications, which forms a foundation for this proposal. The Handbook assesses the potential benefits and costs of energy storage on the national and corporate levels and provides a "technology-neutral," comparative framework utilities can use to formulate application and site-specific assessments of specific technologies. Further insights to the value of DEES will be assessed in this project with Real Options analysis. From the perspective of a regulated T&D utility, the potential added value of such DEES options may serve to justify early DEES deployment that otherwise would not be within their low risk profile. Columbia University has been working on the uses of Real Options for energy management for more than a decade and has recently published the first five parts of a continuing series for the Oil & Gas Journal on the use of Real Options for Lean Energy Management. The study will indicate how and where DEES will likely produce the highest benefits and the potential impacts to rate payers (both participating and non-participating) from Utility R&D deployment of DEES, as well as its on going use as it becomes a commercial time-proven reliable product. Such benefits will be quantified in financial terms plus discussed qualitatively as they relate to the environment, citing flexibility, power security and the overall improved efficiency of the electric grid and supply infrastructure.
This report reflects tireless efforts by hundreds of individuals not identified by name above. They include electrical engineers, information technology experts, and other specialists from across the North American electricity industry, the academic world, regulatory agencies in the U.S. and Canada, the U.S. Department of Energy and its national laboratories, the U.S. Department of Homeland Security, the U.S. Federal Bureau of Investigation, Natural Resources Canada, the Royal Canadian Mounted Police, the Bonneville Power Administration, the Western Area Power Administration, the Tennessee Valley Authority, the North American Electric Reliability Council, PJM Interconnection, Inc., Ontario’s Independent Market Operator, and many other organizations. The members of the U.S.-Canada Power System Outage Task Force thank these individuals, and congratulate them for their dedication and professionalism.
System-wide disturbances that affect many customers across a broad geographic area are rare, but they occur more frequently than a normal distribution of probabilities would predict.
The Distribution Engineering Workstation (DEW) from EDD is being adopted by 12 utilities nationwide as a real-time simulator of a utility's distribution system. DEW maps the topology of the grid, plans peak load leveling, peak shaving, voltage correction, right-sizing of equipment, and automated data modeling. The system is already installed at DTE and portions of the Con Edison distribution grid (Orange and Rockland). Its first use for Transmission was with the Tennessee Valley Authority (TVA) for ERCOT an the Northeastern Interconnect, linking the Transmission and Distribution Grids to simulators at selected Test Sites. DEW is a software package that allows engineering calculations for the distribution system to be linked with those of the transmission system. Using load research statistics or the files accessible through DEW, it becomes an operational tool, managing and recording switching operations. Because it can link all engineering, customer and outage databases it is a database warehouse and through AM/FM systems can produce geographical information system (GIS) maps of the system in its "as planned", "as built" or any state in between. With its new and unique method of calculations (using pointers instead of matrices) it has unprecedented speed of calculation and scope of services. The open architecture allows staff to add data or calculations to the model at any time.
This project represents the convergence of several advanced research and development initiatives which share a common vision for building, evolving, and transforming the current electric delivery infrastructure into what many envision to be the next-generation “Smart Grid”. These different but parallel initiatives include (1) Con Edison’s 3G: System of the Future (3G/SOF) development efforts and (2) Infotility’s GridAgents: DER and Distribution Network Control Framework (GridAgents) currently being developed under the DOE GridWise SBIR program funding. This “Smart Grid” development convergence will continue to be guided by ongoing industry consortium efforts such as GridWise, GridWorks, and Intelligrid with the goal of achieving both ConEdison’s’ and the DOE Office of Electric Delivery and Electric Reliability (OEDER) long term electric delivery goals.
Tokyo Electric Power Company Inc. (TEPCO) is one of the world’s largest privately owned electric power company, which was established in 1951, providing high-quality and stable supply of electricity through its integrated power generation, transmission and distribution system. TEPCO supplies 287TWh to 28 million customers in the Tokyo metropolitan and its surrounding area, which is the political and economic center of Japan. TEPCO’s service area accounts for approximately 40% of the country’s continuously growing economy. EPCO has abundant experience in planning, designing, construction management, operation, and maintenance of its generation, transmission, and distribution systems for large scale redevelopment areas similar to the Hudson Yards Project in the City of New York.
Structure of the Model Weather causes damage (and outages) Wind (regional, gusts, non-linear, trees, water, duration, WC geography) Lightning (frequent or continuous) outages linear in stroke Precipitation (rain, snow, ice) Outages require restoration (resources) Types of damage/outage Restoration takes time (ETR) Available vs. required
Universal One Method to Access All Data Sources Locate User Does Not Need to Know Data Location/Format Grouping Powerful Query Capability, Construct Any Group of Components, Parameters, etc. Views Superimpose Data or View in Different Ways Drill-Down Links to Related Information to Slice and Dice Data Monitor Performance, Availability, Reliability Analysis Statistics, Trend Analysis, and Failure Prediction Automate Automated Data Collection, Analysis, and Status Display Activities Schedule Schedule Activities for Off-Peak Hours Reports Drop Results Into Periodic Status Reports
$370 million in Smart Grid stimulus money is coming to NYC over the next three years How will it change the ways we use electricity? A safer, more secure electric grid? Electric vehicles and a low carbon economy? Lower electricity rates through intelligent energy management in the home and at work?
CO2 and other greenhouse gas emissions from internal combustion engines would drop dramatically The Black Smoke from Diesel vehicle particulates contribute to acid rain, tuberculosis, and other health risks Plug-in Cars require not just Electricity, but also time and locations to recharge the batteries Infrastructure cannot currently provide enough electricity to support this expected influx of EV’s
6+ year Research & Development collaboration Applying sophisticated Machine Learning to the problems of the NYC electric grid Main focus: Preventive Maintenance Predicting what’s likely to break next so it can be fixed before it breaks Also: anomaly detection, capital planning, stray voltage, overtreatment, situational awareness…
Mission: Reliability with Cost Consciousness Prioritize and Optimize the Work Portfolio Significant Con Ed domain expertise with world class R&D from Columbia University Build practical applications with business impact Extend current capabilities - address emergencies Create cutting edge solutions Capture and formalize operating expertise
Goal: to minimize overall curtailment level, while also minimizing the power flows over ratings in transformers and cable sections, over the next several hours; trade-off to be determined through calibration
Necessary for a carbon-free energy system: renewable sources, distributed generation end-use efficiency Improves security and reliability decentralization of an integrated system Secure from physical and cyber threats Self-healing through technology Quality Electricity for the digital economy A digital-age electric system Distributed storage Expanded capacity within the existing footprint
Smart Grid US China Clean Energy Forum Recommendation One Team: Boeing LSI, Con Ed, PG&E, ERCOT, Columbia, (Dr. Chu – HKIT), etc. Scalable Networked Systems M&S, SoS Trades SoSCOE Spiral Development (80/20 Rule) Providing SoSCOE to Utilities Sophisticated Prime for DoD Energy Programs Distributed Generation Energy Storage Energy Efficiency Solar PV Energy GOCO Opportunities
1.0 Introduction This White Paper is submitted to the Department of Energy by a team of Boeing, Columbia University, Con Edison and 5 other electrical power companies to address challenge of smart distribution of power in the urban environment. In this white paper we introduce the concept of treating urban power distribution as a system-of-systems based on a system integration technology that has been developed by the DoD at expenditure of billions of dollars. In a system of systems multiple elements are integrated by a robust information network that provides command and control of the entire entity. The Boeing Company is the world's most experienced system of systems integrator, and will lead the team in the development and prototyping of an urban distribution smart grid. Upon success of the prototype phase the individual members of our team and (it is our expectation) other power companies throughout the country and the world will adopt our results to achieve a revolutionary advance in power system design and operation. 2.0 Our Team Vision for the Urban Smart Grid The Smart Grid is an extremely complex system composed of numerous other systems within it, in other words a system of systems. In an optimized system of systems the individual systems within it are not necessarily each the very best they can be unless their value exceeds their cost. To use a biological analogy, humans are the most successful animals but owls have better eyes, dogs have better hearing, cheetahs have better speed and lions have better strength. Humans are successful because they have superior brains and central nervous systems and one excellent additional system: our hands. Our other systems are adequate. Better eyes, ears, etc. are not present because their added value would be less than their biological cost. Many companies, large and small, are offering innovative solutions for the components and subsystems of the new Smart Grid. Our team endeavors to develop and offer the system integration and the command and control system, i.e. the brain and central nervous system. Our objective is to build upon billions of dollars spent by the DoD to develop system-of-systems technology and apply this to the smart grid. To demonstrate both the need and our capability to perform this role, we have developed and demonstrated a live simulation of a representative highly stressed portion of the grid. Our simulation involved multiple subsystems, operators in the loop, simulated stressful situations and failure modes. The purpose of this white paper is to show how this team is positioned to build upon what we have done to date, incorporate new team members, and expand the simulation to include more of the national grid. Ultimately we would include hardware in the loop and lead the way to a robust, responsive, secure, and economic enhancement of the national electric infrastructure. We would partner with the DOE to determine which would add value in excess of their economic cost. We would accomplish this through simulations that evaluate the performance of the system as a whole in stressing situations of loads and outages. One of the lessons learned in our system of systems projects is that 3.0 Statement of the Technical Problem The economic viability of the United States requires urban electric distribution grids that are reliable, secure, environmentally friendly, and safe for use by all citizens. Building on military and aerospace experience gained over many decades, The Boeing Company realizes that future green urban systems that will rid the United States of hydrocarbon dependence will need Command and Control System of Systems (C2SOS) integration to maintain clean, efficient, reliable, and resilient electric power delivery. The Urban Distribution Smart Grid of the Future must interact continually with massive increases in alternative energy generation and storage owned and operated by many companies and governmental agencies beyond the electric utilities. Yet, the utilities will retain the "responsibility to serve", meaning that their charters require them to be responsible for seeing to it that electric service is maintained (Figure 1).
Enable movement towards a carbon-free energy system: Renewable sources Distributed generation End-use efficiency Improve security and reliability: Secure from physical and cyber threats Self-healing Provide quality electricity for the digital economy: Digital-age electric system Distributed storage Increase efficiency and protect value of the investment in the integrated system Improve investment decision model Permit expanded capacity for growth within the existing footprint
PRESS NOTE SELEX ES, RUDIN MANAGEMENT, AND COLUMBIA UNIVERSITY PARTNER TO DEVELOP OCCUPANCY-DRIVEN DIGITAL BUILDING OPERATING SYSTEM Breakthrough Energy Saving System Debuts at IBCon in June Brewster, NY-XXXX, 2013-Finmeccanica through Selex ES, its global technology systems company, Rudin Management, one of the largest privately held property management companies in New York City, and Columbia University through its School of Engineering and Applied Science have jointly developed a new digital building operating system, a heretofore absent central "Brain" that really makes buildings intelligent. The new system, known as Di-BOSS™, will be unveiled at IBCon 2013 (Intelligent Building Conference) at the Orange County Convention Center in Orlando, June12-13, and will be showcased in Booth #6769. Representatives from all three partners will be on hand to demonstrate the Di-BOSS system with live feeds coming to the convention center floor from Rudin office buildings in New York City. "Di-BOSS is an entirely new kind of digital building operating system solution, a design that could only have resulted from this remarkable, collaborative effort," says Mattia Cavanna, Di-BOSS Project Leader for Finmeccanica/Selex ES. "Columbia's Engineering School provided the primary research and testing expertise, Selex ES brought the Di-BOSS system to fruition, and Rudin operated the system in real time situations to validate and improve its performance capabilities. Our partnership with Columbia and Rudin has proven that the Di-BOSS system can provide building managers and their tenants with significant energy cost savings and enhanced security without sacrificing comfort." Di-BOSS offers several distinguishing innovations. One primary feature is its ability to track occupancy on a large scale. "The technology to link Di-BOSS, the building operating system, with occupancy to control energy use is a cutting edge capability," says Roger Anderson, Di-BOSS Principal Investigator and senior research scientist at Columbia University's Center for Computational Learning Systems. "The Di-BOSS system's forecasting and nowcasting give building managers and operating engineers recommendations to make decisions that significantly improve operating efficiency and better serve the people in the building." The Di-BOSS system is "smart" and uses advanced machine learning algorithms to continuously analyze tenant comfort, energy usage, and other data streams, such as weather forecasts and BMS set points from multiple systems to continuously drive optimal performance. Di-BOSS generates a predictive model that provides lease-required comfort at lower energy consumption rates. The system flags variances between expected and actual performance of all key building systems, and recommends engineers be dispatched to conduct preventive maintenance so that problems are corrected before failures occur. Another Di-BOSS system feature is its ability to simultaneously analyze occupancy and energy consumption trends by tenant. Tenants have visibility through online portals to real time occupancy and energy consumption data for their floors and can see their performance versus other tenants. Since tenants control roughly 60% of a building's energy consumption, Di-BOSS gives tenant facility managers the data and recommendations to affect improvements in their systems that result in real savings for both the tenant and the total building.
The functions that user can perform include: 1. The DO can view real-time load curtailment recommendation and performance metrics comparing to baseline. 2. The DO can change the failure probability and outage duration, customers to be curtailed and the amount of load curtailment, and do a comparative evaluation of this what-if scenario with the real-time and baseline (baseline is curtailing all customers, the current practice), and further to re-run the optimizer to get a new set of load curtailment recommendation. 3. The system administrator can initialize the first time default real-time optimization and performance metrics.
The Command and Control (C2) for the Smart Grid must provide the advanced computational algorithms that will enable distillation of actionable intelligence in a manner that has been demonstrated to be scalable to millions of nodes. Further, it must provide the Secure Smart Grid Common Operating Environment (S2COE) middleware layer capability and cyber-security that is needed to safely and securely manage the complex software applications operating within this System of Systems (SOS) throughout the Smart Grid (new and legacy together), while minimizing costs to the ratepayers. We propose to demonstrate in the New York City and the surrounding region, an integrated C2SOS platform that will provide the intelligence, interoperability among legacy systems as well as new and future systems, and cyber-security that is essential not only for rapid and effective deployment of a Smart Grid in both urban and suburban environs across the country. The nation's power grid has to undergo a massive transformation to be smarter and secure enough to cope with the new demands of distributed generation from solar, wind, tidal and other green sources and equal amounts of distributed storage while achieving reliability, economic, and environmental goals. With information coming from disparate sources, we need a platform for interoperability that is robust and cyber-secure. A key challenge is to make use of the multitude of data from all sources
This project proposal includes integrated development and operation of distributed secondary network load flow models; provides near real-time load profiles for customer locations; validates model load flows from secondary models, provided by installation of new remote devices at strategic customer locations; helps Control Center Operators develop, maintain, and sharpen their situational awareness skills. Additionally, it will improve secondary modeling and load flows to better target grid reinforcement in the networks, minimizing secondary cable failures during peak loading conditions and network outages due to secondary events in the summer. It will also improve the accuracy of the calculated coincident demand for peak summer days. The project will provide a state of the art training tool for the operators for system contingency planning which will help to improve emergency response. This project is a combination of newly proposed projects along with other advancements of ongoing projects that will create additional 14-20 new jobs and cost an estimated $19.0 million. Underground (UG) Distribution Sectionalizing Switches-This project includes installing a combination of automatic and manual sectionalizing overhead switches, to improve the reliability of the overhead distribution systems. The benefits of the project include enhanced reliability by enabling rapid isolation of faulted segments of primary feeders and re-energizing the non-faulted portion of the feeder. It also includes advanced distribution automation and enhances system reliability by creating a more adaptive, integrated/flexible, interactive and optimized grid. The UG Distribution Sectionalizing Switches project is advancement to the existing UG sectionalizing efforts.
• Cleaner More Efficient • Information Reliability • Enterprise Biz Apps • Autonomous Automation • Distributed Energy resources • Demand Side Management • Better safety for Consumers • Integrated Distributed Clean Resources • Less Dependence of foreign Fuels • Electric Vehicle Enablement • Energy Double Star • Distributed Generation from Alternative Energy Resources
Project is to build a scalable smart grid prototype that promotes cyber security, reduces electric demand, increases reliability and energy efficiency, and is cost effective. In addition, the system will enable greater use of renewable energy, other distributed resources, electric vehicle charging and greater consumer participation in the energy mix. The proposed system will be demonstrated in selected sites representing a cross-section of urban and suburban America: the horizontal city (Long Island City, Queens), the vertical city (lower Manhattan), and the suburbs (Orange and Rockland Counties, NY, and northern Bergen County, NJ). New York is America's most populous city as well as home to Wall Street, the Federal Reserve, major medical facilities, and the hub for transAtlantic communication systems and East Coast transportation networks. With one of the highest load densities in the world, it represents the most complex and diverse test bed for urban and suburban Smart Grid challenges in America. As the incubator of the utility industry from Edison's Pearl Street generating station and as pioneer in nuclear power, electric distribution, electric competition and distributed generator interconnection, Con Edison has been an innovator in virtually every major forward movement of the electric grid. The smart grid is at the intersection of sound environmental and economic policy, and New York City (NYC) is already at the leading edge of energy conservation efforts with substantially lower energy use per capita than other national areas. NYC and its environs will also be in the first wave of substantial electric vehicle (EV) markets. Between EVs, renewable energy, and energy efficiency, the smart grid will enable NYC and its suburbs to reduce carbon emissions substantially, and because electricity is cheaper than gasoline, vehicular fuel bills as well. With a secure, interoperable, open smart grid, Con Edison can meet the predicted 30% load increase that widespread EV adoption entails without significant increases to peak electricity requirements. Therefore, the smart grid will result in a cleaner environment and significant savings to area ratepayers and vehicle owners: a triple savings. For the NY and NJ smart grid demonstration project, Con Edison proposes to develop and demonstrate the capability to link the utility and the customer in order to empower intelligent and energy efficient interactions. This will enable consumers to make smart decisions regarding their energy use. The secure, interoperable, open smart grid will enable consumers to charge their EVs with off-peak electricity and power their homes and businesses with wind and solar resources without disrupting the reliability of electricity to their neighbors. The project's goal is to minimize the risks of deploying the smart grid today by providing, in one system of systems, the cyber security, interoperability, scalability, affordability, peak load management, and intelligent, collaborative control required to manage the next generation electric grid in America. To share lessons learned, Con Edison will conduct a Capstone demonstration of smart grid capabilities with others in the energy industry at the end of the project in 2012, as well as through participation in industry standards organizations and professional societies. The proposed smart grid solution will become an economic multiplier, creating thousands of high-technology jobs across the country.
Consensus Strengths and Weaknesses Criterion 1-Strengths: The proposal discusses the incorporation of a community outreach program which utilizes Columbia University and NYU Polytech to provide seminars and technical lectures regarding the smart grid deployment. The utilization of the automated Rules Engine/Dashboard will enable more rapid and less human-directed grid management. This will ultimately be very beneficial as the amount of data being processed by a set of human operators will likely be very complex. Use of the Dashboard is unique. The plan places emphasis on data collection and comparison of results against historic baseline records, cost benefit and operational analysis. The proposed work addresses some goals of the SGD initiative, including the use of advanced digital electronics for use in planning and operations of electric power grid systems. The largest component of the proposed work is the development of general purpose systems engineering software and decision aid software. The adaptation of this software to SG applications has the potential to advance the program metrics. The metric of success is based on data from the current operation of the grid. Selection of NYC may provide a more suitable environment for a smart grid demonstration than other, less populated areas. The reasoning behind this is that NYC contains a sample of most of the infrastructure traditionally found in most communities. Generation (and co-generation) plants including gas and coal are present, as well as renewables, and a variety of building infrastructure from high-priority medical establishments to residential housing, schools, and offices make NYC an ideal setting. The selection of Boeings System of Systems Integration Platform is a wise choice and demonstrates a proven capability to address all major players in the smart grid arena. Integration of a Stochastic Controller is a viable option for providing intelligent control and adaptive response to grid events. This will likely reduce the impact of outages and interruptions to a minimum in addition to increasing efficiency during nominal system operation. The plan contains systematic and periodic clearinghouse reporting. The demonstration sites meet the criteria defined by DOE for proving the benefits of a secure and interoperable open smart grid. Weaknesses:
This Handbook for Assessing Smart Grid Projects was developed for the GridWise Alliance, a diverse group of smart grid stakeholders that includes system operators, utilities, manufacturers, universities, software and communications companies, investors, and consultants. KEMA, Inc. took a lead in writing the handbook with a great deal of input from Alliance members as well as the Edison Electric Institute and their members. This handbook is designed to serve as a reference tool for those organizations and entities that are developing and/or assessing a high quality smart grid project. The handbook provides a legislative background and citations, and lays out a series of metrics that could be considered when developing or assessing a project.
HISTORY AND WHAT WE ARE BRINGING TO THE TABLE ON DAY 1: 6 YRS OF DEVELOPMENT IN COMMAND AND CONTROL AT CONED, ADP ASC PATENT, AND KEEPER OF THE IMPROVEMENT METRICS ABOVE BASELINE
As part of the Edison Project at Con Edison, we have built the first two tiers of intelligent control for the electric grid: process control and engineering anticipation. We have Process Controls in place in all Con Edison Control Centers that provide machine learning for real time contingency analysis and capital asset prioritization that are used to predict intelligent maintenance actions, and then we complete the automated scoring of the actual outcomes of the work to validate and improve the preventive maintenance program of Con Edison. No other utility that we are aware of has such an operational system in place.
This study was conducted from September through November 2009 as part of the Utility Energy Research led by Smart Grid operation executive and Stanford University Graduate School of Business and Master of Science Electrical Engineer alumni. We completed our interviews with over 100 people from different functional roles at 55 Utilities comprised of Investor-Owned Utilities (IOU), Municipalities and Electric Cooperatives across North America (Figure 1). The Utilities interviewed are distributed in 18 U.S. states (Figure 2) and one Canadian province with coverage areas extending into 10 adjacent states. Of the interview participants, 42 percent are holding operational leadership positions with an average of 26 years in their role; remaining 58 percent is comprised of an equal distribution of C-Level management (CEO, COO and CIO) and technical (AMI specialist, communication specialist, power engineers) functions and project managers.
2030 Smart Grid vision Unleasing the power of consumer response and distributed resources The emergence of the “prosumer” and the ability of the individual to direct their energy world Achieving new levels conservation, reliability and efficiency 2020 - Role of smart grid and Con Edison Optimized network leveraging distributed and central station resources Resulting in enhanced Reliability Cost savings Environmental Impact Innovation New level of grid utilization and maintenance – self healing operations and condition based maintenance Realization of the City as the most energy efficient environment 2010 - Demonstration role Smart Grid requires algorithms and control systems to deal with the flood of information in a near real time Prove new technology and avoid new sources of failure Modularity for a very integrated energy system Security Balance centralized control versus Customer control Architecture for data management & warehouse costs
Command and Control In partnership with Boeing, Columbia University, and The Prosser Group, Con Edison proposes to design and deploy intelligent network centric command and control system-of-systems (C2SOS) in conjunction with demand management, distributed generation, and energy efficiency projects. This project will provide real time situational awareness and transparency via an Integrated System Model of the electric transmission grid enables targeted management and intervention to resolve issues as they arise. Accommodated effective, plug and play compatibility amongst new, green technologies that have the potential to disrupt grid function. This is a new project that will create 90 to 100 new jobs and cost an estimated $61.7 million. Due to additional funding from the associated partners listed above, Con Edison is only asking for 25% (approximately $15.4 million) funding for the project.
In NYC and northern NJ, C2SOS will also demonstrate the control of energy, storage hardware in key Smart Grid demonstration sites to provide needed stability, peak load reduction, and integration with intermittent renewable energy resources, particularly for emergency response." • "In particular, the system of systems approach carries risk of integration failure as multiple systems from different vendors come together to form a whole. Experience with large system of systems projects provides several mitigation strategies to minimize this risk of integration failure."
Commitment at the highest levels ofColumbia University’s Administration Energy Engineer dedicated to working the program Cooperation and teamwork attitude toward supporting the goals of the program Financial resources and consultant support
We have developed a Machine Learning Energy Tuning System (MLS) for owners who manage a portfolio of large energy consuming facilities such as microgrids (e.g., Columbia's Morningside Heights and Manhattanville campuses, Luke Air Force Base), commercial and residential skyscrapers (Rudin Management in Manhattan), electric vehicle fleets (Fedex Express and UPS) and large construction projects (Turner Construction, Lend Lease). The MLS is patented and is being deployed as part of a DOE Smart Grid Demonstration Project at Con Edison in New York City (U.S. Patent 7,395,252 and others pending. In order to tune successfully, the MLS must contain modules for 1) data acquisition of the energy consumption and environmental footprint of each individual building within the portfolio, 2) automate metrics and measure against a baseline of improvement that must be determined in advance, 3). Use machine learning models such as Support Vector Machines and Martingale Boosting to continually learn from the past performance of each building in the facility, so that 4) tuning of the energy consumption can be optimized using Approximate Dynamic Programming (ADP) that learns from building history and advanced weather diagnostics to 5) predict future energy consumption for all buildings within the portfolio, so that 6) cost benefit can be verified and 7) a feedback loop established so that real option economic models can maintain the facility on the efficient frontier of energy efficiency (figure below). Autonomous Control Systems for large buildings and facilities are more difficult than those required to control indoor and site-specific systems (e.g. factory assembly lines, petrochemical plants, and nuclear power plants). Below we describe the Columbia Adaptive Stochastic Control (ASC) system for load and source management in real-time of large facilities operations. Electricity, steam, and water usage in large facilities is dominated by stochastic (statistical) variability, primarily driven by the vagaries of the weather. Advanced dynamic control is required for simultaneous management of real time pricing, curtailable loads, Electric Vehicle recharging, solar, wind and other distributed generation sources, along with many potential forms of energy storage such as batteries, ice, and compressed air.
In FY 2007, nearly 24% of the DoD energy cost of approximately $15B was associated with facilities. Energy load management and alternative source initiatives have significant impacts as energy savings can be redirected within the DoD budget and reduce DoD demand on local electrical grids. With over 750M sq.ft. of facility space on 90 installations, the US Air Force has instituted corporate level initiatives linked with energy industry leaders to provide guidance and incentives targeting the highest consuming spaces. At installations such as Luke AFB, AZ, these initiatives have improved efficiency, facility control, load control and source management measures on an individual level by addressing the 12 facilities that comprise the top 50 percent of identified energy consumption. While these facility level measures have had a significant impact with positive cost-benefit and savings-to-investment ratios, a significant challenge exists in identifying and capitalizing on the opportunity for capturing the remainder of potential savings. Accomplishing this requires an innovative and tightly integrated approach as these loads are often widely distributed. At Luke, for instance, they are distributed over 375 facilities that vary in size, construction standards and class of use and are largely unmetered. Achieving facility-level feedback sufficient to support advanced management concepts cannot be achieved using such conventional approaches as facility and sub-facility metering alone. By developing and deploying a fully integrated suite of energy measurement, management and optimization tools, this project will create a solution for identifying and capturing distributed opportunities for energy savings and lead to greater ROI in decision making for military installation energy issues. 5. Technology Description a) Technical Objective: Implement a low-cost, minimally invasive, installation-level energy management system to support operational and strategic decisions. : b) Technology Description and c) Maturity Levels. The ASU approach is a seamlessly integrated, end-to-end approach that connects facilities within the installation into a system-of-systems level management loop using proven technologies. The project deploys a cost effective combination of direct and indirect metering applications to close the gap on information for energy consumption sources and facility load levels. With the information gap closed, a real-time, web-based monitoring system supports a
Smart buildings and cities are evolving the application of a broad range of simple and sophisticated technologies to observe, sense, and collect data; various types of sensors and visual data (still image and video). This data collection represents 7 by 24 coverage throughout the year under a range of contextual framing (e.g., seasonal, holiday, etc…) to examine and analytically learn from such examples as: 1. Dynamic Power usage on the utility grid. 2. Operational and environmental control in buildings (e.g., heating/air conditioning, elevators, etc… 3. What are the dynamic metrics of different manufacturing and delivery facilities in a big city.. 4. How do we address power utilization optimally across these multiple building types collectively? 5. How do we promote the conversion to Electric Vehicles in large urban cities, particularly delivery trucks and taxis?
Smart Building solution will address problems important not just to Rudin – highly relevant for broader marketplace
Allows FNM to recruit the right leadership and resources team (external and internal) Establishes FNM-wide governance Retains operational risk within the operating companies Complies with federal US government requirements (e.g., industrial security and cost accounting) Consolidates common and FNM-wide tasks in one shared organization
We have developed a Machine Learning Energy Control System (MLS) for owners who manage a portfolio of large energy consuming facilities such as microgrids (e.g., Columbia's Morningside Heights and Manhattanville campuses, Luke Air Force Base), commercial and residential skyscrapers (Rudin Management in Manhattan), electric vehicle fleets (Fedex Express and UPS) and large construction projects (Turner Construction, Lend Lease). The MLS is patented and is being deployed as part of a DOE Smart Grid Demonstration Project at Con Edison in New York City (U.S. Patent 7,395,252 and others pending. In order to tune successfully, the MLS must contain modules for 1) data acquisition of the energy consumption and environmental footprint of each individual building within the portfolio, 2) automate metrics and measure against a baseline of improvement that must be determined in advance, 3). Use machine learning models such as Support Vector Machines and Martingale Boosting to continually learn from the past performance of each building in the facility, so that 4) Control of the energy consumption can be optimized using Approximate Dynamic Programming (ADP) that learns from building history and advanced weather diagnostics to 5) predict future energy consumption for all buildings within the portfolio, so that 6) cost benefit can be verified and 7) a feedback loop established so that real option economic models can maintain the facility on the efficient frontier of energy efficiency (figure below). Autonomous Control Systems for large buildings and facilities are more difficult than those required to control indoor and site-specific systems (e.g. factory assembly lines, petrochemical plants, and nuclear power plants). Below we describe an Adaptive Stochastic Control (ASC) system for load and source management in real-time of large facilities operations. Electricity, steam, and water usage in large facilities is dominated by stochastic (statistical) variability, primarily driven by the vagaries of the weather. Advanced dynamic control is required for simultaneous management of real time pricing, curtailable loads, Electric Vehicle recharging, solar, wind and other distributed generation sources, along with many potential forms of energy storage such as batteries, ice, and compressed air.
Columbia University’s Center for Computational Learning Systems (CCLS) and our development partner, the Castle Laboratory of Princeton University, will install a Columbia Machine Learning (ML) System, based on its version 1.0 that is optimizing energy consumption of six commercial office buildings in Manhattan (Figure 1). This AZ DOD ML system will be a website that will integrate with information from the ASU “Campus Metabolism” control system on DOD bases. We will analyze the historical record of weather patterns and electricity, gas, and water consumption at the bases and predict optimal tuning of the energy management systems. DOD will provide access to their energy profiles, past and future, for hourly-and-less demand and consumption data for electricity, gas and water (Figure 2). ASU will provide look-ahead weather forecasting, and historical actuals specific to the DOD bases to be used to train the Machine Learning System. Columbia will provide actionable energy management decisions via messaging to the Campus Metabolism site to suggest recommended timing of demand to be matched to HVAC and other energy systems of each building, based on risk assessment using our newly deployed Machine Learning System that will be adapted as required for this project. We will then establish a feedback loop to feed the Machine Learning system with results from the Campus Metabolism system tracking actual consumption. The combined systems should continuously improve the efficiency of the bases’ total energy consumption.
The use of Distributed Electric Energy Storage (DEES) for the real time support and optimization of the electric generation, transmission and distribution (G,T&D) system has been limited to date to pumped hydro, primarily due to a lack of cost-effective options and/or sufficient value bases, as well as actual field experience. Recent developments in advanced energy storage technology, including a number of demonstration and commercial projects, are providing new opportunities to use energy storage in grid stabilization, grid operation support, distribution power quality, and load shifting applications. Our team proposes to characterize the leading DEES markets for New York State, including a projection of the respective capacities and range of values. We will then drill down into the detailed cost, benefit, risk, and uncertainty benefits for use of an exemplary DEES technology, a 10 MW sodium-sulfur (NAS) super-battery in an urban substation in a critical Load Pocket of New York City.
Story 1: Real-time information, no what-if scenarios Operator reviews the values on the objectives achieved by the LSOC for the current list of recommended curtailments and makes his/her decision. Operator clicks on any of the customers on the list and the selected customer is shown on map.
The Metrics web application is to measure base line and compare performance metrics on the Con Ed power system. It is also able to measure performance Columbia contributions to the smart grid implementation. The functions include GridView dashboard and intelligent charts. In Metrics, performance metrics are plotted versus time and against other attributes of demonstration networks and Con Ed power system.
Key Innovation: ID evolving patterns over time and recommend repair of chronic anomalies in the primary and secondary by prioritizing field actions based on predicted benefits in Mean Time Between Failure –– before they become significant outages
Key Innovation: ID evolving patterns over time and recommend repair of chronic anomalies in the primary and secondary by prioritizing field actions based on predicted benefits in Mean Time Between Failure –– before they become significant outages.
After Power losses from Hurricanes, the replacement solution is not "putting back overhead power lines exactly as before the storm", but absorbing the higher cost one time of putthing the power cables underground like every other major utility.
In short, NERC’s investigation concludes that: • Several entities violated NERC operating policies and planning standards, and those violations contributed directly to the start of the cascading blackout. • The existing process for monitoring and ensuring compliance with NERC and regional reliability standards was inadequate to identify and resolve specific compliance violations before those violations led to a cascading blackout. • Reliability coordinators and control areas have adopted differing interpretations of the functions, responsibilities, authorities, and capabilities needed to operate a reliable power system. • Problems identified in studies of prior large-scale blackouts were repeated, including deficiencies in vegetation management, operator training, and tools to help operators properly visualize system conditions. • In some regions, data used to model loads and generators were inaccurate due to a lack of verification with actual system data and field-testing. • Planning studies, design assumptions, and facilities ratings were not consistently shared and were not subject to adequate peer review. • Available system protection technologies were not consistently applied to optimize the ability to slow or stop an uncontrolled cascading failure of the power system. • Communications between system operators were not effective and hampered their ability to recognize the developing system emergency
Enable movement towards a carbon-free energy system: Renewable sources Distributed generation End-use efficiency Improve security and reliability: Secure from physical and cyber threats Self-healing Provide quality electricity for the digital economy: Digital-age electric system Distributed storage Protect value of the investment in the integrated system Permit expanded capacity for growth within the existing footprint
A key finding of the blackout investigators is that violations of existing NERC reliability standards contributed directly to the blackout. Pending enactment of federal reliability legislation creating a framework for enforcement of mandatory reliability standards, and with the encouragement of the Stakeholders Committee, the board is determined to obtain full compliance with all existing and future reliability standards and intends to use all legitimate means available to achieve that end. The board therefore resolves to: • Receive specific information on all violations of NERC standards, including the identities of the parties involved; • Take firm actions to improve compliance with NERC reliability standards; • Provide greater transparency to violations of standards, while respecting the confidential nature of some information and the need for a fair and deliberate due process; and • Inform and work closely with the Federal Energy Regulatory Commission and other applicable federal, state, and provincial regulatory authorities in the United States, Canada, and Mexico as needed to ensure public interests are met with respect to compliance with reliability standards. The board expresses its appreciation to the blackout investigators and the Steering Group for their objective and thorough work in preparing a report of recommended NERC actions. With a few clarifications, the board approves the report and directs implementation of the recommended actions. The board holds the assigned committees and organizations accountable to report to the board the progress in completing the recommended actions, and intends itself to publicly report those results. The board recognizes the possibility that this action plan may have to be adapted as additional analysis is completed, but stresses the need to move forward immediately with the actions as stated.
The Aug. 14, 2003, blackout in the northeastern United States was not an isolated event. Slow response times of mechanical switches, lack of automated analysis of problems, and an inability to see the whole grid in real time are contributing to a noticeable increase in failures of the present electric grid. These problems have caused a noticeable increase in blackouts and brownouts since 1998.
This issue of Power & Energy features two articles that examine the power grid and how to prevent another regional blackout. Lawrence T. Papay looks at the near term—how we can increase grid reliability through the introduction of new technology. Roger Anderson and Albert Boulanger, setting their sights farther down the line, advocate creating an advanced system architecture that would automatically recognize and respond to disruptions on the grid. As Papay, Anderson, and Boulanger remind us, efforts to defend the grid against more regional outages is really a race against time. If we don't make a concerted, coordinated effort to shore up the system's reliability, last summer's blackout may become a mild prelude to an onrushing disaster.
The use of Distributed Electric Energy Storage (DEES) for the real time support and optimization of the electric generation, transmission and distribution (GT&D) system has been limited to date to pumped hydro, primarily due to a lack of cost-effective options and/or sufficient value bases, as well as actual field experience. Recent developments in advanced energy storage technology, including a number of demonstration and commercial projects, are providing new opportunities to use energy storage in grid stabilization, grid operation support, distribution power quality, and load shifting applications.
We recognize that development of a Smart Grid must include computer sciences, geosciences, public policy, law, business, electrical, mechanical, and nuclear and earth engineering, and successful research will require collaborations both inside and outside the university. Lamont-Doherty Earth Observatory, through the efforts of Roger Anderson and Albert Boulanger, has already made a substantial investment in this effort and has laid the groundwork for a major University-wide research program. They have created an informal “Energy Forum” consisting of SIPA, SEAS, Law, the American Museum of Natural History and individuals responsible for operating Columbia’s Powerhouse. In addition to organizing the efforts at Columbia, the principals of the Smart Grid effort have assembled a broader NewEnergy NorthEast (NE2) team consisting of Brookhaven National Laboratory, the Center for Economic and Environmental Partnership, Inc. (CEEP), New York University and Con Edison. Lamont has established joint visiting appointments between Roger Anderson, Albert Boulanger, and Robert Hall (Head, Infrastructure Technologies) at Brookhaven National Laboratories for the explicit purposes of jointly pursuing the NE2 Smart Grid research goals described in this proposal.
The Smart Grid will first be created virtually, through a computer modeling and simulation testbed of the New York state, city and NPCC grid, and then, in several New York State test-beds that are Hybrids -- new hardware tests separate from the grid system but tied to the virtual testbed software through power electronics that can simulate connectivity to the grid. We will demonstrate the utility and effectiveness of a wide range of new technologies managed by the new-generation computer control systems, first in computer models, then in the Hybrid test-beds, and finally in controlled demonstration projects on the live grid.
This chapter explains the major events—electri- cal, computer, and human—that occurred as the blackout evolved on August 14, 2003, and identi- fies the causes of the initiation of the blackout.
Members of the U.S.-Canada Power System Outage Task Force and Its Three Working Groups
Short, localized outages occur on power systems fairly frequently. System-wide disturbances that affect many customers across a broad geographic area are rare, but they occur more frequently than a normal distribution of probabilities would pre- dict.
The Clean Energy Advocates (CEA) appreciate the opportunity to comment on the future direction of the New York System Benefits Charge (SBC). Our diverse coalition of public interest environmental organizations, consumer advocates, public health interests, clean energy technology manufacturers, laborers, trade associations, energy services providers, and green marketers is emblematic of the many constituencies who have been well-served by the SBC, and of the abiding interest within our communities to see the program fulfill the extent of its promise in delivering economic, environmental and energy security benefits to New York State. In these comments, we will both answer the questions about the future of the SBC posed by the Commission in its Notice dated January 28, 2005, but will also offer our affirmative vision of how the SBC program should continue, expand and grow in strength and effectiveness in the future. As is discussed below, CEA strongly support the SBC program, and urge that it be extended and expanded in several important respects, including the addition of a significant natural gas efficiency program.
This is a response to Public Service Commission (PSC) Staff request for comments on 14 questions concerning the future of the System Benefit Charge. I offer the comments on behalf of Center International Earth Science Information Network (CIESIN), Earth Institute, Columbia University, and in consultation with Roger Anderson, Doherty Senior Scientist and Albert Boulanger, Senior Staff Associate, of the Lamont-Doherty Earth Observatory, Earth Institute, David Waltz, Director, Center for Computational Learning, School of Engineering and Applied Science, Columbia University and Chris Jones and Robert Yaro at the Regional Plan Association.
This project will constitute a transformational step towards a city–wide infrastructure control system across all critical energy infrastructures that will most significantly impact resiliency, efficiency, and reliability in meeting the ever-growing demands of great urban centers in environmentally sound ways.
There are 12 major Load Pockets in New York City, within which it is difficult to balance supply and demand. The problem will only get worse as demand grows and generation diversifies. Solution: Design a combination of a Smarter Grid, in-city generation from NYPA and others, DG & DS, Automated Cap Bank Balancing, & Real-Time Phasor Monitoring to eliminate or better control the Load Pockets
The future power systems have to be flexible and reliable enough to integrate the future breakthrough in the energy sector. The Consortium for an Electrical Infrastructure to support Digital Society (CEIDS) design a vision shared by the entire energy value chain to depict these attribute of future power systems1. A way to achieve this is to design an auto-adaptative grid, a “self-healing” grid which would include a very high levels of monitoring and automation. To guarantee proper operation of the entire system, all of the automation have to rely on powerful monitoring, simulation and modeling tools. A key performance factor for these predictive analysis tools is its capability to analyze the system faster-than-real-time, proposing control actions to the operator for ensuring system security.
NYM-Infraguard.us is a group of volunteers from private infrastructure industries that works with the FBI to secure critical infrastructure in the Metropolitan New York City area. From drinking water supplies to communications systems, chemical production processes to agricultural resources, and all energy infrastructure, Americans depend on a select group of critical infrastructures to sustain our way of life. Any attempts to harm or destroy these resources would directly impact the security of the United States and its citizens. Most of these systems and services are owned and operated by private industry. Therefore, the protection of our nation’s infrastructure cannot be accomplished by the federal government alone. It requires coordinated action from numerous stakeholders – including government, the private sector, law enforcement, academia and concerned citizens such as members of Infraguard.
A webcast discussion exploring the best ideas for improving American infrastructure and building a better, safer future. To view the webcast, visit www.nsf.gov/bridges APRIL 10, 2008 e Smart Grid 12:30 p.m. ET Second-by-second information sharing among households, utilities and even individual appliances may revolutionize the grids that distribute electricity throughout the country. Panelists will explore how to make the grid more resilient and nimble, saving energy and forestalling blackouts. Panelists to include Roger N Anderson, Columbia University.
To understand national response to outage in regional networks, need to consider two interrelated issues: 1. power transfer in a particular direction impacts line flows in large portion of system -- this impact is called the Power Transfer Distribution Factor (PTDF) 2. Once a line is congested, any new power transfers above 5% more can not take place on the congested line! Outage on a single line can constrain congestion in many different directions Once a line is congested, any generators that would increase loading on the congested line are prevented from selling into that market.
General Motors Corporation and Suzuki Motor Corporation announced today a new agreement to collaborate in the development of fuel cell vehicles. The agreement will focus on developing small car applications for fuel cell technology, as part of long term cooperation on future product development programs.
Wind generation capacity in the USA has reached 4,685 MW (27 states)* California 1822 MW Texas 1095 MW Iowa 423 MW 10% increase in 2002 alone
CAES is a peaking gas turbine power plant that consumes less than 40% of the gas used in conventional gas turbine to produce the same amount of electric output power. This is because, unlike conventional gas turbines that consume about 2/3 of their input fuel to compress air at the time of generation, CAES pre-compresses air using the low cost electricity from the power grid at off-peak times and utilizes that energy later along with some gas fuel to generate electricity as needed. The compressed air is often stored in appropriate underground mines or caverns created inside salt rocks. It takes about 1.5 to 2 years to create such a cavern by dissolving salt. Deployment Status:
Hawaii has the highest per unit electricity costs in the United States. In spite of the state’s abundance of natural resources, the islands have no fossil fuel reserves, and therefore are forced to depend on imported fuels to generate the vast majority of their electricity. Unable to draw power from neighboring states, each island must generate sufficient production to meet their economy’s growing demands. Hawaii’s economy is dependent on selling itself as a tropical paradise; therefore it is crucial that the state’s energy production not be a source of major pollution.
Power grid/marketplace overview How energy users impact that system Various energy efficiency methods & technologies, including status of alternate energy systems Opinions/projections/prognostications Suggestions for entering the energy biz
Dr. Roger N. Anderson and Albert Boulanger from the Lamont- Doherty Earth Observatory, a member of The Earth Institute at Columbia University, along with colleagues from Rice University's Center for Nano Scale Science &Technology, the Texas Energy Center, and the Texas Superconductivity Center, have developed the framework for a "Smart Electric Grid," and plans are underway to test their system in Texas as well as the northeastern U.S.
What is Machine Learning (ML)? General idea A brief history Supervised learning methods SVM (Support Vector Machine) Boosting Con Edison example Reinforcement Learning Real Options Mining example Messages Power of ML Wide range of practical ML applications